flair-bio/amplify-ablations
Updated
None defined yet.
Welcome to FLAIR Lab. We build and study language models for proteins, genomes, and transcriptomes. Our goal is to accelerate drug discovery by making these models more efficient and capable. Our work spans the full pipeline, from data curation and architecture design to pre-training, fine-tuning, alignment, evaluation, and interpretability.
FLAIR Lab is based at Mila – Quebec AI Institute and the Université de Montréal (DIRO), and is led by Quentin Fournier.
We share our models and datasets here on Hugging Face. All releases are available under the MIT License.